Geophysical Journal International

Geophys. J. Int. (2015) 201, 528–542 doi: 10.1093/gji/ggv041 GJI Geodynamics and tectonics

Fault activity in the epicentral area of the 1580 Dover Strait (Pas-de-Calais) earthquake (northwestern Europe)

D. Garc´ıa-Moreno,1,2 K. Verbeeck,1 T. Camelbeeck,1 M. De Batist,2 F. Oggioni,3 O. Zurita Hurtado,2 W. Versteeg,2,4 H. Jomard,5 J. S. Collier,3 S. Gupta,3 A. Trentesaux6 and K. Vanneste1 1Royal Observatory of Belgium, Av. Circulaire 3, 1180 Uccle, Brussels, Belgium. E-mail: [email protected] 2Renard Centre of Marine Geology, Ghent University. Krijgslaan 281 S.8,B-9000 Ghent, Belgium 3Department of Earth Science and Engineering, Imperial College London. Prince Consort Road, London SW72BP,United Kingdom 4Vlaams Instituut voor De Zee. InnovOcean site, Wandelaarkaai 7,B-8400 Oostende, Belgium 5Institut de Radioprotection et de Suretˆ eNucl´ eaire.´ Av. de la Division Leclerc 31, Fontenay-aux-Roses, France

6Universite´ Lille 1,Lab.Geosyst´ emes` UMR 8217 Lille1/CNRS, 59 655 Villeneuve d’Ascq, France Downloaded from

Accepted 2015 January 20. Received 2015 January 19; in original form 2014 February 12

SUMMARY http://gji.oxfordjournals.org/ On 1580 April 6 one of the most destructive earthquakes of northwestern Europe took place in the Dover Strait (Pas de Calais). The epicentre of this seismic event, the magnitude of which is estimated to have been about 6.0, has been located in the offshore continuation of the North Artois shear zone, a major Variscan tectonic structure that traverses the Dover Strait. The location of this and two other moderate magnitude historical earthquakes in the Dover Strait suggests that the North Artois shear zone or some of its fault segments may be

presently active. In order to investigate the possible fault activity in the epicentral area of at North Dakota State University on May 20, 2015 the AD 1580 earthquake, we have gathered a large set of bathymetric and seismic-reflection data covering the almost-entire width of the Dover Strait. These data have revealed a broad structural zone comprising several subparallel WNW–ESE trending faults and folds, some of them significantly offsetting the Cretaceous bedrock. The geophysical investigation has also shown some indication of possible Quaternary fault activity. However, this activity only appears to have affected the lowermost layers of the sediment infilling Middle Pleistocene palaeobasins. This indicates that, if these faults have been active since Middle Pleistocene, their slip rates must have been very low. Hence, the AD 1580 earthquake appears to be a very infrequent event in the Dover Strait, representing a good example of the moderate magnitude earthquakes that sometimes occur in plate interiors on faults with unknown historical seismicity. Key words: Palaeoseismology; Intraplate processes; Submarine tectonics; Neotectonics; Europe.

faults and hence the possible present-day seismic hazard related to 1 INTRODUCTION them. Assessing the seismic hazard associated with plate-interior tectonic This study focuses on the poorly known offshore Sangatte Fault, structures is generally a very difficult task. For instance, the seis- which traverses the marine Dover Strait (Pas de Calais) from micity in these areas is generally too low to be studied by means of Sangatte (northeastern France) to Folkestone (southeastern Eng- classic seismotectonic approaches. In addition, destructive earth- land; Figs 1 and 2). This fault is part of the North Artois shear zone quakes are rare and they usually appear randomly distributed in (Fig. 1) and it is believed to be the probable source of the magnitude space and are sometimes located in places with no known histor- ∼6.0 earthquake that occurred offshore in AD 1580 (Camelbeeck ical seismicity (Zoback & Grollimund 2001; Camelbeeck et al. et al. 2007). Historical information suggests that this earthquake 2007). This apparent randomness of the seismicity might be the produced damage equivalent to MSK intensity VIII in Calais and consequence of recurrence intervals too long to have produced two Dover and about VI in London, and it was felt as far as Koln¨ significant ruptures of the same feature in historical times (Stein & (Germany) to the east and York (England) to the north (Neilson Mazzotti 2007). It is thus necessary to adopt palaeoseismic meth- et al. 1984; Melville et al. 1996; Musson 2004). According to ods in order to understand and characterise the activity of intraplate the empirical relationships of Wells & Coppersmith (1994), the

528 C The Authors 2015. Published by Oxford University Press on behalf of The Royal Astronomical Society. Fault activity in the Dover Strait 529 inferred magnitude of the AD 1580 earthquake would correspond different reactivations of structures inherited from the Variscan to an average fault length of 10 km and to a slip of 0.25–0.30 m. Orogeny (e.g. Chadwick et al. 1993; Van Vliet-Lanoe¨ et al. 1998, However, nothing is presently known about any active fault located 2002a,b;Mansyet al. 2003; Minguely et al. 2010). Since its for- in the Dover Strait. mation, the North Artois shear zone has passed from the original Apart from the AD 1580 earthquake, the Sangatte fault has only compressional setting (Variscan Orogeny) through extension related been associated with two other seismic events of moderate magni- to the opening of the Tethyan and Atlantic ocean basins during tude. These are the MS ∼ 4.0 historical earthquake that occurred Jurassic and Cretaceous times, and again to compression during the in 1776 (Melville et al. 1996) and the Mw 4.0 Folkestone earth- tectonic inversion that started in the early Palaeogene epoch due to quake of 2007 (Ottemoller¨ et al. 2009). The hypocentre of the latter the Alpine Orogeny (Vandycke et al. 1988; Bergerat & Vandycke was instrumentally localized about 5 ± 2 km below Folkestone 1994; Van Vliet-Lanoe¨ et al. 2002a,b, 2004;Mansyet al. 2003; (Ottemoller¨ et al. 2009). Minguely et al. 2010). The presence of major infrastructures and the densely populated From an analysis of gravity data, Everaerts & Mansy (2001) cities within the area of influence of an earthquake similar to the concluded that the major fault segments comprising the North Ar- AD 1580 event make it very important to assess the nature of the tois shear zone have lengths ranging from 15 to 40 km and are Sangatte Fault and its tectonic activity over time. In order to do so, arranged as right-stepping en-echelon fault zones (Fig. 2). Based we have gathered a large set of seismic-reflection and bathymetric on their gravity map, Camelbeeck et al. (2007) distinguished five data from the epicentral area of this earthquake. In this paper, we major faults: that is Marqueffles Fault, Ruitz Fault, Pernes Fault, will present and discuss the interpretation of these geophysical data, Landrethun Fault and Sangatte Fault (see Fig. 2). Not only their ac- Downloaded from which resulted in the first high-resolution bathymetric map of the tivity, but the precise geometry of these faults is still poorly known. Dover Strait and provided a clear visualization of the main tectonic This is especially true for the mostly submarine Sangatte Fault, on structures and their activity over time. which there has been no specific study published until now; despite the fact that it corresponds to the most likely geological structure

capable of having generated the AD 1580 event (e.g. Camelbeeck http://gji.oxfordjournals.org/ 2 TECTONIC AND GEOLOGICAL et al. 2007). SETTINGS The Quaternary activity of the North Artois shear zone is still The North Artois shear zone is a complex fault-and-fold system debated, as no conclusive field evidence of recent tectonic deforma- defining the western part of the Variscan Midi–Eifel Thrust Front tion has yet been associated with any of its fault segments. Indirect (Fig. 1), which separates the Ardenne Massif and Paris Basin to the evidence includes possible extensional faults identified in the San- south (hanging wall) from the London–Brabant Massif to the north gatte cliff (Van Vliet-Lanoe¨ et al. 2000) and minor right-lateral (footwall). The present-day geometry of this fault zone results from deformations affecting river development and Quaternary fluvial several phases of post-Palaeozoic deformation that have induced and aeolian deposits in northeastern France (Colbeaux et al. 1981; at North Dakota State University on May 20, 2015

Figure 1. Onshore geological map of northwestern Europe updated with the geology of the English Channel and major tectonic structures (De Bethune´ & Bouckaert 1968). Known historical and instrumental earthquakes (Mw ≥ 3) are indicated scaled by magnitude according to catalogues from the Royal Observatory of Belgium and Grunthal¨ et al. (2009). 530 D. Garc´ıa-Moreno et al. Downloaded from http://gji.oxfordjournals.org/

Figure 2. Interpreted horizontal derivative of the Bouguer gravity anomaly (see Everaerts & Mansy 2001; Camelbeeck et al. 2007). The Bouguer anomaly was calculated using the gravity database of the Royal Observatory of Belgium. For France and United Kingdom the data were provided by the French and British Geological Surveys respectively. All gravity data are referenced to the gravity datum Uccle (1976) (IGSN71–0.048mGal). The density reduction used for the at North Dakota State University on May 20, 2015 calculation of the Bouguer anomaly on land was 2.67. Above the sea, the free-air anomaly was used. Grey coloured circle: epicentre of the 2007 Folkestone earthquake (Ottemoller¨ et al. 2009); dashed ellipsoid: isoseismal indicating MSK Intensity VII–VIII during the 1580 earthquake (Melville et al. 1996); LF, Landrethun Fault; SF, Sangatte fault; RF, Ruitz fault; PF, Pernes fault and MF, Marqueffles fault.

Van Vliet-Lanoe¨ et al. 2002a;Mansyet al. 2003). The latter suggest long due to the dynamic conditions this area has been subjected to right-lateral strike-slip deformation, which is in good agreement during the Holocene, with strong sediment reworking, erosion and with the NNW–SSE orientation of the maximum horizontal stress sediment starvation (e.g. Hamblin et al. 1992). On the other hand, if measured near Boulogne by Froidevaux et al. (1980). However, it the Sangatte Fault reaches the surface and has ruptured several times disagrees with the direction of the deformation suggested by the in earthquakes similar to the AD 1580 since the Fosse Dangeard focal mechanism calculated for the 2007 Folkestone earthquake and the Lobourg Channel were formed, the cumulated deformation (Ottemoller¨ et al. 2009), which indicates a left-lateral rupture along produced on these erosional features should be measurable. a WNW–SES fault (see Ottemoller¨ et al. 2009). The Fosse Dangeard is a kilometre-scale network of sediment- According to some authors, the Sangatte Fault may have been filled palaeobasins that are incised several tens of metres into in active extension during the deposition of the Gault Clay and bedrock (e.g. Destombes et al. 1975; Smith 1985; James et al. 2002). Lower Chalk formations in the Cretaceous (Warren & Harris 1996; The origin of these buried palaeobasins is presently uncertain, al- Ottemoller¨ et al. 2009; Minguely et al. 2010). On the 1:250 000 though most authors agree that they were probably formed during offshore geological maps published in 1988 and 1989 by the British the Middle Pleistocene. Initially, they were attributed to glacial pro- Geological Survey (BGS) and the compilation of geological data cesses (Destombes et al. 1975). Nevertheless, most of the recent performed by James et al. (2002), the Sangatte Fault is represented studies seem to agree that none of the Pleistocene ice sheets ex- as several WNW–ESE trending minor faults and folds traversing tended that far south (e.g. Clark et al. 2004). Hamblin et al. (1992) the Cretaceous bedrock (Fig. 3). This structure also traverses two thus proposed that the Fosse Dangeard may be the result of the com- significant Quaternary features situated in the centre of the Dover bination of fluvial erosion and tidal scouring during, respectively, Strait (Figs 3 and 4): a broad palaeochannel known as the Lobourg Pleistocene low and high sea level stands. Van Vliet-Lanoe¨ et al. Channel and a complex sediment-filled palaeobasin network known (2004) suggested a possible tectonic origin for these basins. How- as the ‘Fosse Dangeard’ (see Destombes et al. 1975; James et al. ever, considering the amount of deformation necessary to create 2002). Elsewhere Quaternary sedimentary features are mainly lim- such large basins, similar offsets should be observable in Quater- ited to a number of major (stable) and minor (mobile) Holocene nary deposits along the on-shore continuation of these structures, sandbanks (Mellet et al. 2013) separated by areas with thin or ab- which is not the case (e.g. Minguely et al. 2010). A tectonic origin sent sedimentary cover (see James et al. 2002). Recent tectonic of the fosse is thus very unlikely. Presently, it is widely accepted movements affecting the latter features will not be preserved for that these palaeobasins have been originally carved into the bedrock Fault activity in the Dover Strait 531 Downloaded from http://gji.oxfordjournals.org/

Figure 3. Bedrock geology of the Dover Strait/Pas-de-Calais area constrained by combining the 1:250 000 geological maps published by the BGS in 1988 and 1989 (sheets 51N00 and 50N00) and the maps published by James et al. (2002). White rectangle indicates the area shown in Fig. 12. The yellow line represents at North Dakota State University on May 20, 2015 the location of the Fosse Dangeard according to James et al. (2002).

Figure 4. Merged bathymetry, single- (dark blue lines) and multichannel (white lines) seismic-reflection profiles gathered for this study, and boreholes archived at BRGM. Hb92: location of the single-channel seismic-reflection profile interpreted by Hamblin et al. (1992). Projection for this and following figures: UTM-31N (WSG84). 532 D. Garc´ıa-Moreno et al. by a mega-flood produced during the breaching of the Weald-Artois part of the Dover Strait by the RCMG and Lille University in 2002 ridge, which used to separate the southern North Sea from the En- on board of RV Sepia II (Dangeard I Cruise), as well as with parts glish Channel (Smith 1985; Gupta et al. 2007). of two bathymetric datasets provided by Imperial College London. Dating of the Fosse Dangeard is currently limited to pollen ex- The seismic dataset mainly consists of parallel lines. This was tracted from a 50 m borehole collected from one of its sediment- due to limitations of collecting lines across the Dover Strait due to filled palaeobasins (Destombes et al. 1975). This dating suggests strong tides and shipping lane restrictions. The high quality bathy- that these sediments were deposited during the Brørup interstadial of metric data was therefore invaluable to correlate between the seismic the Weichselian glaciation. Based on this dating, Destombes et al. profiles. Notably, the high resolution of this (horizontal resolution (1975) argued that the Fosse Dangeard should have been formed 1.5–5 m) and the single-channel seismic-reflection (vertical resolu- in a previous stage, as the features themselves reach depths up to tion 1–3 m) datasets permitted the visualization of features as small 100 m. In addition, the sedimentary infill of these palaeobasins as 1.5 m in the centre of the strait. Specific technical details on presents several erosional surfaces (Destombes et al. 1975), attest- the vertical and horizontal resolutions of the geophysical data, as ing to different phases of scouring and infilling. These authors thus well as on their acquisition and processing, can be consulted in the proposed that the Fosse Dangeard was formed and firstly filled-up electronic supplement attached to this paper. with sediments during the previous (Saalian) glaciation (0.35–0.13 Ma). This feature may actually be much older, as considering the mega-flood hypothesis postulated by Smith (1985) and Gupta et al. (2007) its formation would be associated with the opening of the Downloaded from 4 INTERPRETATION OF THE Dover Strait. The exact age of this event is presently unknown. How- GEOPHYSICAL DATA ever, it is generally accepted (e.g. Gibbard 1995; Toucanne et al. 2009) that it occurred during the Elsterian/Anglian glacial stage 4.1 Seismic stratigraphic interpretation (0.48–0.40 Ma).

The Lobourg Channel is a broad NE–SW palaeochannel extend- In this study, correlations between geology and seismic stratigraphy http://gji.oxfordjournals.org/ ing along the Dover Strait and cut predominantly into bedrock were first done in the northwesternmost seismic-reflection profile (James et al. 2002). This palaeochannel extends to the south- collected for this study and correlated afterwards southeastward west into a complex anastomosing system of valleys that can be from one profile to another. The interpretation of the first profile mapped continuously from a few kilometres northeastward of the is based on the equivalence established by Hamblin et al. (1992) Dover Strait to the western approaches (e.g. Smith 1985;Lerico- between geology and the seismic stratigraphy of one single-channel lais et al. 2003; Gupta et al. 2007). Presently, two main hypotheses seismic-reflection profile acquired nearby the English shore (see are proposed to explain this major erosional feature. Some au- location in Fig. 4). Wecomplemented this interpretation with several thors (e.g. Gibbard 1995; Lericolais et al. 2003) suggest that these interpreted boreholes (Fig. 4) stored at the BRGM (www.brgm.fr) palaeovalleys were part of an enormous palaeodrainage system and the available geological maps from the Dover Strait (Fig. 3). at North Dakota State University on May 20, 2015 (Channel/Manche River) that was fed by the North European rivers Correlations between seismic profiles were performed using the few during the major Middle and Late Pleistocene marine low stands. available cross-lines and the merged bathymetry. The latter proved Others authors have proposed that this feature may have been cre- invaluable in correlating between the seismic profiles, as many of the ated by several episodes of Middle Pleistocene catastrophic flooding sedimentary formations form distinct geomorphological features on (e.g. Smith 1985; Gupta et al. 2007). In any case, both groups of au- the seafloor that can be followed along the entire width of the strait thors agree that the Lobourg Channel formed following the breach (see Figs 5, 6 and 8). The geological interpretation presented in of the Weald-Artois ridge. This palaeochannel was probably ac- this study is limited to the first 100 m below the seabed, as we tive during all the major Middle and Upper Pleistocene marine low were not able to completely remove the first seabed multiple in stands, being finally submerged at the end of the last glaciation— the multichannel seismic-reflection data. Hence, we have very little beginning of the Holocene epoch (e.g. Gibbard 1995; Lericolais information on the structure below the Wealden Beds of Lower et al. 2003; Toucanne et al. 2009). Cretaceous age (Fig. 7). The strata imaged in the seismic-reflection profiles correspond to Lower and Upper Cretaceous sedimentary rocks and the sediments infilling Quaternary palaeobasins and palaeochannels or forming 3 METHODOLOGY AND AVAILABLE sandbanks. Palaeogene and Neogene units are missing in this area. DATA The Quaternary palaeobasins and palaeochannels are the easiest This study is based on the interpretation of several sets of bathymet- features to identify in the seismic-reflection data due to the distinct ric and seismic-reflection data (Fig. 4) collected across the gravity unconformity formed by their basal erosional surfaces (BES). The anomaly observed by Everaerts & Mansy (2001) traversing the seismic facies infilling the largest features (i.e. the Fosse Dangeard) Dover Strait (see Fig. 2). The principal seismic-reflection dataset comprise several subfacies composed of seismic reflections with comes from three geophysical campaigns organized by the Royal low to moderate amplitude. These subfacies are separated (and Observatory of Belgium (ROB) in collaboration with the Renard truncated) by thin packages of high amplitude reflections that seems Centre of Marine Geology (RCMG) in 2010 and 2012 on board of to correspond to erosional surfaces carved into the sedimentary infill RV Belgica (cruise reports 2010/09, 2012/03 and 2012/25 available (Fig. 9). from www.mumm.ac.be). These campaigns consisted of the acqui- During this study, we mapped the various Quaternary deposits, sition of 17 multichannel (total 309 line km) and 48 single-channel constructing isopach maps along the different seismic-reflection (total 487 line km) seismic-reflection profiles over the entire width profiles by subtracting the two-way times (TWT) of the BES and of the strait and more than 300 km2 of multibeam bathymetric data seabed and assuming a velocity of 1750 m s−1, as proposed by from its central part. These data were complemented with 20 other Arthur et al. (1997) from log tests performed for the Channel Tunnel high-resolution seismic-reflection profiles acquired from the central geo-engineering investigations. We have included the isopach map Fault activity in the Dover Strait 533 Downloaded from http://gji.oxfordjournals.org/

Figure 5. Structural map derived from the interpretation of the seismic reflection and multibeam bathymetric datasets collected for this study. Major bathymetric features are also indicated. Ant-Syn, Anticline/Syncline system. GR, Greensand Ridge. at North Dakota State University on May 20, 2015

Figure 6. Structural interpretation and isopach map of the sediments infilling Quaternary palaeobasins and palaeochannels. (a) Enlarged view of the purple square indicated in the structural map. Location of the single-channel seismic-reflection profile shown in Fig. 7(a) is indicated. 534 D. Garc´ıa-Moreno et al. Downloaded from http://gji.oxfordjournals.org/

Figure 7. (a) Interpreted single-channel seismic-reflection profile (see Fig. 6 for location). (b) Enlarged view of the rectangle (dash outline) indicated in (a). P.C.: Wealden buried palaeochannels; pale violet lines in (b): seismic horizons selected to illustrate the strata geometry around F3. The depth conversion in this and the following interpreted seismic profiles is based on a mean constant velocity of 2000 m s–1. in Fig. 6 to illustrate the distribution of the Quaternary deposits with (3) The marine deposits known as Gault Clay (see Hamblin et al. regards to the main tectonic structures. 1992; Hopson et al. 2010) present two distinct parts. The upper Concerning the seismic-stratigraphy of the Cretaceous bedrock, half of this formation is acoustically almost transparent, whereas we identified the following seismic units, from Lower to Upper the lower one has a more heterogeneous acoustic character with the at North Dakota State University on May 20, 2015 Cretaceous age: appearance of low amplitude reflections (Fig. 7). We believe that these two subunits are equivalent to the Upper and Lower Gault (1) The upper part of the fluvial and lacustrine Wealden Beds Clay described by Owen (1975) and Woods et al. (1995). This (see Hamblin et al. 1992; Radley & Allen 2012), which are char- subdivision is clearly observed near the British coast. However, it acterised by a succession of well-defined low and high amplitude becomes more ambiguous towards France, where a seismic unit reflections in its upper part that become discontinuous and heteroge- showing very similar characteristics to the lower Gault Clay lies neous with depth. The seismic-reflection profiles show significant directly on top of a seismic unit resembling the Sandgate Beds lateral variability of this unit’s seismic facies, with the presence (Figs 8 and 9). The Folkestone Beds appear thus to be missing in of buried palaeochannels and other erosional/depositional features that area. consistent with its continental origin (Fig. 7b). The thickness of the Gault Clay formation ranges between 30 (2) The shallow marine/shoreline deposits known as the Lower and 50 m across the survey area, generally getting thinner to the Greensand (see Ruffell & Wach 1991; Hamblin et al. 1992; Hopson southeast. This is consistent with its outcrop in northern France, et al. 2010), which are divided into four units in the eastern English where only 11 m of this unit are recorded (e.g. Hamblin et al. Channel (e.g. Hamblin et al. 1992; Hopson et al. 2010); from old to 1992). young: Atherfield Clay, Hythe Beds, Sandgate Beds and Folkestone (4) Finally, the marine Chalk Group, which is generally subdi- Beds. The Atherfield Clay presents an almost completely acous- vided in three units (Hamblin et al. 1992; Mortimore 2011), i.e. the tically transparent facies (Fig. 7a). The Hythe Beds appear as a Lower, Middle and Upper Chalk formations. The Lower Chalk for- package of reflections of moderate amplitude. The Sandgate Beds mation is especially easy to identify in the seismic-reflection profiles consist of seismic subunits composed of reflections of moderate thanks to the double diffraction produced by the Channel Tunnel amplitude alternating with seismic subunits comprising reflections (Fig. 8), which was bored into it (Warren & Harris 1996). Moreover, with relatively lower amplitudes (Fig. 7a). Finally, the Folkestone the lower boundary of the Lower Chalk formation is highlighted by Beds are characterised by very diffuse and wavy reflections present- the contrast between the reflections of moderate amplitude charac- ing hyperbolic diffractions (Fig. 7a). In general, the Lower Green- terising its seismic facies and the almost acoustically transparent sand formation thins to the southeast. This is especially true for the Upper Gault Clay that lies underneath (Fig. 7a). The boundaries be- upper part of the sequence, with the Sandgate Beds thinning by as tween the three Chalk formations themselves are much less evident much as 25 per cent and the Folkestone Beds being no longer rec- because their seismic facies vary significantly from one profile to ognizable in the southeastern half of the strait (Fig. 8). Assuming another. In this study, we have therefore mapped these boundaries that the Folkestone Beds are absent in the southeast (see next point), by combining our seismic-reflection and bathymetric datasets with the Lower Greensand formation is ∼30 m thicker offshore England the BRGM borehole descriptions and the bedrock-geological map than offshore France (compare Figs 7 and 8). showninFig.3. Fault activity in the Dover Strait 535 Downloaded from http://gji.oxfordjournals.org/ at North Dakota State University on May 20, 2015

Figure 8. 3-D block showing the central and southeastern merged bathymetry in relationship with the interpretation of one of the single-channel seismic- reflection profiles. Coloured circles in the bathymetry represent the fault traces and monocline axis inferred from the seismic investigation. CT: hyperbolic diffraction due to the Channel tunnel; Fig. 9a: location of the fault scarp shown in Fig. 9 (a). Inset: structural map with the area shown in the 3-D block (dash line) indicated. Blue line: location of the seismic-reflection profile. See Fig. 6 for the Inset depth colour scale, isopach map shade scale and structural interpretation.

All of the sedimentary formations described above outcrop at the 4.2 Tectonic interpretation seafloor, where they define several subparallel WNW–ESE ridges The geometry of the Cretaceous strata outcropping in the Dover and sharp slopes resulting from differential erosion (Figs 5, 6 and Strait is strongly controlled by several WNW–ESE subparallel mi- 8). Especially marked in the bathymetry are some of the units com- nor and major synclines, anticlines and fault systems (Figs 5–12). posing the Lower Greensand and Chalk Groups, which withstand We can distinguish two deformation styles: (1) a broad asymmetric erosion better than the Gault Clay and Wealden Beds formations. anticline/syncline system, a north-facing monocline structure and a For instance, the Hythe Beds form a prominent ridge (in this paper reverse fault (F1), all three accommodating apparent compressional referred to as Greensand ridge) on the seafloor that can be fol- deformation; and (2) several faults accommodating vertical offsets lowed over almost the entire width of the Dover Strait (see Figs 8 more typical of extensional or strike-slip settings. and 9). The different response to erosion of the Cretaceous sed- The north-facing monocline structure traverses the almost-entire imentary formations outcropping at the seafloor has also empha- width of the Dover Strait with its axis coinciding on the seafloor sized some old tectonic-related features, like minor folds and fault with the ESE–WNW striking Greensand ridge (Fig. 8). This struc- scarps (Figs 6aand8). In addition to the outcropping bedrock, the ture is without doubt the most significant deformational feature seafloor presents two prominent Quaternary erosional/sedimentary located in the Dover Strait, in which the otherwise subhorizontal features: the aforementioned broad NE–SW palaeochannel, known Cretaceous strata are significantly tilted to the east (Fig. 8). The as the Lobourg Channel, and two kilometre-scale, elongated, sub- monocline structure accommodates most of the deformation docu- parallel sandbanks, called The Varne and The Colbart (Fig. 4). The mented by the seismic investigation. In fact, the vertical offset pro- latter two are associated with postglacial Holocene sea level rise duced by this structure is obscured in the seismic-reflection profiles (Reynaud et al. 2003; Mellet et al. 2013). In addition to these major by the first seismic multiple. This means that it must be larger than sedimentary/erosional features, the bathymetry shows several mi- the maximum vertical penetration of the single-channel seismic- nor Quaternary sandbanks and erosional features (troughs, grooves, reflection data. We can thus assume offsets greater than 100 m channels, etc.) carved in the seafloor over the entire study area (see (see Fig. 8). Figs 5, 6 and 8). 536 D. Garc´ıa-Moreno et al. Downloaded from http://gji.oxfordjournals.org/ at North Dakota State University on May 20, 2015 Figure 9. Selected parts of three single-channel seismic-reflection profiles (dark pink lines in structural map) acquired across fault F1. G. Ridge, Greensand ridge; BES, basal erosional surface; ES, erosional surface; MF, minor fault (only identified in this profile). See Figs 6 and 7 for isopach map shade scale and geological interpretation of the seismic-reflection profiles, respectively. Yellow and red lines in the structural map indicate the location of the seismic-reflection profiles shown in Figs 10(a) & (b) and 11(a)–(d).

F1 is located in the central part of the strait, running parallel are seen farther to the southeast at the intersection of the Quaternary to the monocline structure over more than 13 km (Figs 5 and 6). palaeobasin and F1 (Figs 9b and c). These features are also limited This high-angle (60–80◦) west-dipping fault juxtaposes the Gault to the BES and lowermost strata infilling the palaeobasin and do not Clay and Lower Chalk formations against the lower units of the show any predominant deformation style. Lower Greensand (Figs 9b and c), suggesting significant reverse Finally, the anticline/syncline system is located ∼6kmtothe offsets. This fault rarely reaches the seafloor due to the presence of northeast of the monocline/F1 structure presenting a similar trend a sediment-filled palaeobasin carved along its almost-entire length and lateral extent as F1 (see Figs 5–7). Besides some minor non- (Fig. 9). The presence of this palaeobasin and the high position deformed Quaternary palaeochannels carved into this structure, de- of the seismic multiple prevented the direct measurement of the posits younger than Upper Cretaceous have been eroded or not vertical offset induced in the Cretaceous bedrock by this structure. deposited on top of it, preventing the assessment of the complete A rough estimation of 40–50 m can be calculated though assuming activity history of this feature. that the thickness of the Lower Greensand formation does not vary The second set of tectonic structures, that is those accommodating too much between the seismic-reflection profiles shown in Figs 8 apparent extensional and/or strike-slip deformations, is composed and 9 (c) (distance ∼5 km). of several minor and major high-angle (60–90◦)—fault systems. F1 seems to terminate to the northwest at one of the Fosse The most significant are: the WNW–ESE trending faults F3 and F4 Dangeard palaeobasins (Fig. 9a). At that location, the BES and and the W–E trending faults F5, F6, F7 and F8 (see Figs 5–12). lower internal strata of this palaeobasin are ∼5 m offset to the Faults F3, F4 and F5 show similar geometries and characteristics, southwest (Fig. 9a). This suggests normal or strike-slip faulting in- which suggests that they are probably different segments of the same stead of the reverse deformation observed along F1 farther to the system. All three faults dip to the southwest and offset normally southeast (Fig. 9c). This might indicate, on the one hand, that F1 the Wealden Beds. The offsets produced by these faults vary both accommodated some normal or/and strike-slip deformation follow- laterally and with depth, ranging between 5 and 40 m. This is well ing the formation of the Fosse Dangeard. On the other hand, it is evidenced in Fig. 11, where we observed an increase of the offset also possible that we are looking at two different faults present- from northwest to southeast from ∼9to∼24 m in less than 2 km ing opposite deformational style. In any case, the offset disappears along strike. The offset decreases again to ∼5 m 3.5 km further above the first erosional surface that truncates the lower layers of the to the southeast (Fig. 11d). We have also noticed that the Wealden palaeobasin infill (Fig. 9a). Other troughs and small buried scarps strata located to the southwest (hanging wall) of F5 is generally Fault activity in the Dover Strait 537 Downloaded from http://gji.oxfordjournals.org/

Figure 10. Parts of two single-channel seismic-reflection profiles (location indicated in Fig. 9) traversing faults F4 and F6, and a major palaeobasin of the Fosse Dangeard. Orange dashed line in (b): palaeobasin BES. at North Dakota State University on May 20, 2015 downwarped near the fault plane (Fig. 11). This downward bending of the strata varies significantly along strike and with depth, being sometimes more pronounced in the upper seismic units than in the lower ones (Fig. 11b). The fault system F3–F4–F5 presents minor scarps in the outcrop- ping Wealden Beds (Fig. 6a). However, it does not seem to reach the seabed surface through the Quaternary sediment-filled palaeobasins and palaeochannels. Rather, the possible Quaternary deformations accommodated by these faults are limited to the BES and lowermost infill of the palaeobasins (see Figs 7band10b). Only F5 seems to extend locally all the way to the seafloor along the western slope of a partially filled depression carved into the Lobourg Channel (Figs 8 and 11a). Nevertheless, despite the clear throws observed at depth in seismic-reflection profiles acquired less than 200 m apart, the scarp associated with F5 is restricted to the partially filled depres- sion (Figs 8 and 11). No other scarp or offset have been observed within the Lobourg Channel or any of the other sediment-filled palaeobasins traversed by this fault system (see Figs 7b, 10band 11). In fact, the deformation possibly linked to the tectonic activity of the system F3–F4–F5 at other locations consists of small changes in the geometry of the palaeobasin’s BES and lower infill near to or across the fault planes (e.g. Fig. 7b). This suggests that F5 was exhumed at that location by a concentration of the scouring rather than by fault activity. Figure 11. Selected parts of four single-channel seismic-reflection profiles Faults F6 and F7 are both high angle E–W trending faults acquired across F5 (location indicated in Fig. 9). Red lines in (a), (b), (c) separated by a right step-over of 1.4 km (Fig. 5). These faults and (d): offsets induced in the Wealden strata by F5; orange dash line: present localized minor scarps on the seafloor within outcrop- Quaternary palaeobasins BES. ping Wealden Beds and in locations with thin Quaternary sedi- mentary cover (Fig. 8). The scarps are not continuous through the Lobourg Channel, where the erosional/depositional features located 538 D. Garc´ıa-Moreno et al.

20 m (Fig. 12). No Quaternary deformation has been observed in the available geophysical data associated with this structure. Apart from the tectonic structures described above, several other faults and folds are present in our study area deforming the Creta- ceous bedrock (see Figs 5–11). We do not describe these structures in detail because they are either of minor importance (e.g. F2) with respect to the ones described above or poorly imaged in the seismic investigation (e.g. F3b). However, these structures are included in the structural and geological maps shown in Figs 5 and 12.

5 DISCUSSION The new bathymetric data gathered for this study have revealed a seafloor morphology strongly influenced by the erosional and sediment-starved conditions presently prevailing in the Dover Strait. Indeed, the geophysical data shows that, apart from some minor and major sandbanks, there has been almost no Holocene deposition in Downloaded from this area. This environment has resulted in very good preservation of the major Late-Quaternary erosional features (e.g. the Lobourg Channel) and a range of the Cretaceous sedimentary formations and older tectonic structures outcropping at the seafloor (see Figs 5–12).

The major Quaternary sediment-filled palaeobasins (i.e. Fosse http://gji.oxfordjournals.org/ Dangeard) are located along or in between the main tectonic struc- tures identified in this area. However, even though some of them present possible minor deformations affecting their BES, neither their morphology nor the locations of their depocentres are con- trolled by the deformations accommodated by the tectonic struc- tures traversing or nearby them (Figs 6, 9 and 10). These features were thus formed by erosion as already stated by most of the studies previously performed in this area (e.g. Destombes et al. 1975; Smith 1985; Hamblin et al. 1992; Gupta et al. 2007). at North Dakota State University on May 20, 2015 Figure 12. Selected parts of the single-channel seismic-reflection profile The tectonic structures identified in the Dover Strait are mainly acquired across F8. Red line indicates the offset induced in the Wealden deforming Lower Cretaceous sedimentary formations. The largest strata by F8; orange dash line: Quaternary palaeobasins and palaeochannels deformations are concentrated on the WNW–ESE north-facing BES. monocline and fault F1. These two structures present the same apparent polarity (i.e. F1 reverse offset is in agreement with the in it appear unaffected by these faults (Fig. 8). F6 seems to be re- monocline folding), suggesting that their associated deformations lated to a minor step or buried scarp disrupting the southwestern were produced during a similar tectonic regime. Their proximity slope of the palaeobasin located underneath the Lobourg channel and the continuity of the Greensand ridge (which is the surface (Fig. 10b). However, the low amplitude of the reflections infilling expression of both the monocline and hanging wall of F1) indicate this palaeobasin prevented the characterisation of this possible de- that these two deformational features are probably different expres- formation. In any case, it does not seem to extend beyond the thin sions of the same structure. In other words, the monocline is most package of high amplitude reflections corresponding to an erosional likely a blind reverse fault system, whose master fault or one of surface carved into the sedimentary infill (see Fig. 10b). its associated segments may reach the surface in the central part The sense and amount of vertical deformation induced by faults of the strait as F1. The monocline/F1 structure, together with the F6 and F7 in the Cretaceous bedrock is also uncertain. We were anticline/syncline system identified to its northeast, thus point to a unable to estimate either of these parameters due to the seismic compressional tectonic episode during which the whole Cretaceous multiple and the strong differences between the seismic facies at sequence was folded and faulted. This episode most likely corre- either side of the fault planes (Figs 8 and 10a). It is actually possi- sponds to the tectonic inversion that started in early Palaeogene ble that this fault system has accommodated significant horizontal times (e.g. Mansy et al. 2003). offset, which would explain our inability to correlate the seismic The offset produced by the monocline/F1 structure in the Cre- reflections across the fault. Indeed, the Greensand ridge (axis of the taceous strata has almost no expression in the seafloor (Fig. 5), monocline) appears to have been right-laterally bent about 350 m indicating that more than 100 m of bedrock have been eroded from where it intersects the inferred continuation of F7 (Fig. 5). Unfortu- this area since these structures were formed. This has resulted in the nately, we did not find any other right-laterally deformed bathymet- exhumation of the Wealden Beds to the south of the north-facing ric markers to positively link this bend with activity on the fault. monocline (Figs 8 and 9). Younger Cretaceous formations outcrop Finally, F8 is only imaged in one of the single-channel seismic- only northwestward of this structure. Notably, the Lower Greensand reflection profiles (Fig. 12) and two of the multichannel seismic is restricted to the monocline axis in the central part of the strait. This lines, preventing an accurate analysis of its activity. This nearly observation disagrees with the bedrock-geological map presently W–E trending southwest-dipping fault appears to offset normally available, which proposed a broader outcrop for the Lower Green- the lower Cretaceous and probably upper Jurassic strata by about sand formation (see Fig. 3). Neither does that bedrock-geological Fault activity in the Dover Strait 539 Downloaded from http://gji.oxfordjournals.org/ Figure 13. Bedrock geology of the Dover Strait derived from the interpretation of the seismic-reflection and multibeam bathymetric datasets available for this study. map indicate the significant reverse offset produced by F1 in the during shorter periods of time than the Wealden Beds. Hence, their central part of the strait. Fig. 13 shows the new bedrock-geological original normal offsets would have been much smaller and, so, over- map derived from our investigation, including the aforementioned compensated by the Tertiary tectonic inversion. This is consistent observations and the lithology and thickness variations of the Lower with the observations of Minguely et al. (2010) and Underhill & Greensand and Gault Clay formations described in Section 4.1. Paterson (1998) in the North Artois shear zone (northeastern France) The compressional deformation produced by the monocline/F1 and the Wessex Basin (Southern England), respectively. Both stud- at North Dakota State University on May 20, 2015 structure does not seem to have influenced the shape of the ies showed several examples of faults that offset the upper and Quaternary palaeobasins located in this area. On the contrary, the younger sedimentary strata reversely, while the offset in the lower Quaternary offset associated with the possible northwestern con- and older strata remains normal. tinuation of F1 follows the sense of its dip, resulting in geometries Alternatively, the particular characteristics of the fault systems more typical of extensional faults (Fig. 9a). That suggests that the F3–F4–F5 and F6–F7 and the apparent change of polarity along F1 compressional regime ended prior to the formation of the Fosse could be due to strike-slip deformation. This interpretation would Dangeard. be in good agreement with the high dipping angles of F6 and F7 Significant deformation of the Lower Cretaceous strata has also fault planes and the apparent right-lateral bend of the Greensand been observed along fault systems F3–F4–F5, F6–F7 and F8, most ridge observed at its intersection with the possible prolongation of them associated with apparent normal and/or strike-slip faulting. of F7 (Fig. 5). The right-lateral offset of the monocline structure A characteristic of these faults is the significant lateral variation of would actually imply that this possible strike-slip motion took place their associated offsets and the vertical and horizontal variations following the Tertiary compressional episode. If that was so, it is of the bending of the strata near their fault planes (see Figs 7, 10 plausible that the system F3–F4–F5 may also have accommodated and 11). Similar variations of the offset and seismic-strata geome- some strike-slip deformation following the tectonic inversion. How- tries have been observed in the central English Channel by Collier ever, the observed up-sequence variations in dip of the seismic-strata et al. (2006). These authors interpreted that as the consequence near the system F3–F4–F5 fault planes are better explained by the of the tectonic inversion of initially syn-sedimentary extensional normal-faulting hypothesis. faults combined with the heterogeneous sediment composition and The limited penetration of the seismic-reflection profiles available spatial distribution of the Lower Cretaceous continental deposits. for this study prevented imaging the basement and the relationship This interpretation is consistent with the lateral and vertical vari- among the different faults at depth. It is remarkable though that all ation of the offsets and bending of the strata observed along the major structures are located within the gravity anomaly interpreted system F3–F4–F5. The fact that we do not observe an inversion of as the Sangatte Fault (compare Figs 2 and 5), suggesting that they the original normal offsets associated with these structures is be- all belong to the same structure: the Sangatte Fault system. cause they only traverse Wealden and older sedimentary formations We have not been able to identify the actual source of the AD in our study area. That is, the reverse deformation produced in the 1580 or the possible deformation associated with the rupture that Wealden Beds during the Tertiary tectonic inversion may not have caused this earthquake. However, we have identified possible Qua- been enough to compensate the significant pre-existing extensional ternary fault activity from irregularities, offsets and changes of offsets. On the other hand, the monocline/F1 and anticline/syncline the dip within the lower sedimentary infill and BES of the Pleis- systems are deforming much younger sediments (Aptian–Turonian tocene palaeobasins (Figs 7–11). Especially notable are the irregular age), which were subjected to the Cretaceous extensional settings cross-sectional morphologies of some minor palaeobasins located 540 D. Garc´ıa-Moreno et al. along F3 (Fig. 7b) and the apparent ∼5 m vertical offset of one compressional monocline structure. However, we have not been able of the palaeobasin BES by the possible northwestern continuation to determine when this possible strike-slip deformation started or of F1 (Fig. 9a). Nevertheless, neither of these deformations seems whether it has been ongoing during the Quaternary. to extend above the first erosional surface that truncates the lower The possible Quaternary tectonic activity of the main structures layers of the palaeobasins sedimentary infill (see Figs 7b, 9 and identified in this study seems to have been very low in recent times. 10b). As a matter of fact, these faults have virtually no expres- Whilst the lack of recent deformation in the bathymetry could be sion on the seafloor where their paths cross the various Quaternary explained by the strong erosional tidal conditions in the Dover Strait, palaeobasins and palaeochannels. This means that the deformation the fact that we do not see it in the upper layers of the palaeobasins cumulated since the lower infill was eroded must be below the max- sedimentary infill suggests that, if this exists, it must be below the imum resolution of the geophysical data, which is 1–3 m for the resolution (1–3 m) of the single-channel seismic-reflection data. The seismic-reflection profiles and 1.5 m for the bathymetric data. cumulated deformation accommodated since Middle Pleistocene by The Dover Strait was exposed to subaerial erosional/depositional this fault system may actually be lower than 5 m. Therefore, if the settings during the last three major Pleistocene glacial stages (e.g. Sangatte Fault system has been active during the late-Quaternary Gibbard 1995). Consequently, we cannot rule out the possibility period, the recurrence interval of large earthquakes must have been that the erosional surfaces that are unaffected by the faults were very long. Thus, earthquakes similar to the AD 1580 event have formed during the last glacial stage (0.11–0.012 Ma). This is also been rare in this area. Earthquakes of greater magnitudes seem to true for the apparent lack of deformation in the seafloor. Indeed, have been even rarer or non-existent. Nonetheless, the size of the any minor deformations could have been washed out during the last structures identified in this study suggests that they could potentially Downloaded from marine low stand; especially those affecting the Lobourg Channel, rupture in earthquakes with magnitudes up to 7. which may have confined a significant river during the last glacial In conclusion, this study supports the episodic character of the maximum (0.03–0.025 Ma) (e.g. Gibbard 1995). Therefore, with seismic activity typical of tectonic structures located in plate inte- the data presently available we can only state that the Sangatte riors. These structures present ‘short’ periods of activity followed

Fault system has produced deformations less than 1–3 m since the by long periods of seismic quiescence that may correspond with http://gji.oxfordjournals.org/ Holocene started. The cumulated offset since the Fosse Dangeard the activity of nearby regional structures. This ‘migration’ of the was formed in Middle Pleistocene appears to be of ∼5m.However, seismicity is well observed in northwestern Europe, where most of this could be smaller, as the buried scarp on which we measured the known moderate and large historical earthquakes have occurred this offset could have been caused or exaggerated by the erosional at different locations (Camelbeeck et al. 2007). process that originated the Fosse Dangeard. In summary, earthquakes of magnitudes equal to or higher than 6 seem to have been very rare in this area since Middle Pleistocene. ACKNOWLEDGEMENTS This makes the AD 1580 an exceptional event. As a matter of fact,

We would like to thank Koen De Rycker for the great job he did at North Dakota State University on May 20, 2015 assuming that the cumulated deformation accommodated by the during the data acquisition. Many thanks also to Tine Missiaen, Sangatte Fault system during the Holocene epoch is just below the Rindert Janssens, Matthias Baeye, Philipp Kempf, Geoffroy Van maximal resolution of the geophysical data (1.5 m), we can estimate Kriekinge, Pierre Legrand, Alexander Comeyne, Naomi De Roeck, by applying the equations of Wells & Coppersmith (1994) that this Sophie Van de Putte and Billie Heene for their assistance during deformation would correspond either to the occurrence of a single data acquisition. Thanks go as well to NIOZ for renting us the M ∼7.0 earthquake produced by a fault with an average length of w multichannel acquisition equipment, and in particular to Hendrik ∼40 km, or to very few earthquakes with magnitudes equal to or De Haas and Leon Michael Wuis for their hard work during the lower than 6.0 on shorter faults. More importantly, these estimations acquisition of the data. We would also like to thank the Captains suggest that none of the earthquakes produced by the Sangatte Fault and Crews of the RV Belgica and R/V INSU ‘Cotesˆ de la Manche’ system since that epoch has exceeded a maximum magnitude of for their excellent work during the geophysical campaigns. Ship 7.0, which is consistent with the length of the individual structures Time on the RV Belgica was provided by BELSPO and RBINS–OD composing it. Nature, and managed by MUMM (www.mumm.ac.be). Le Service Hydrographique et Oceanographique´ de la Marine (SHOM) pro- vided single-beam data for French waters (Contract No. 205/2011) 6 CONCLUSIONS and the UK Hydrographic Office and UK Maritime and Coastguard Agency provided single and multibeam data for UK waters. This study demonstrates the presence of a more complex yet con- tinuous fault-and-fold system traversing the Dover Strait/Pas-de- Calais than previously published. This fault-and-fold system off- sets the Cretaceous sediment strata enough to influence the bedrock REFERENCES geology outcropping at the seabed. The system is composed of Arthur, J.C.R., Phillips, G. & McCormick, C.R., 1997. High definition seis- several major and minor tectonic structures accommodating either mic for Channel Tunnel marine route, in Modern Geophysics in Engineer- compressional deformation or deformations more typical of exten- ing Geology, Vol. 12, pp. 327–334, eds McCann, D.M., Eddleston, M., sional and/or strike-slip regimes. The comparison between these Fenning, P.J. & Reeves, G.M., Geological Society Engineering Geology, different styles of deformation and the known regional tectonic his- Special Publication. Bergerat, F. & Vandycke, S., 1994. Paleostress analysis and geodynamical tory suggests that the compressional structures were formed during implications of Cretaceous-Tertiary faulting in Kent and the Boulonnais, the Tertiary compressional phase. Apparent normal deformations J. Geol. Soc., 151, 439–448. would be then more likely linked to the Jurassic–Cretaceous ex- Camelbeeck, T. et al., 2007. Relevance of active faulting and seismicity tensional phase and/or to a strike-slip episode that took place after studies to assessments of long-term earthquake activity and maximum the Tertiary tectonic inversion. The latter seems to be the case for magnitude in intraplate northwest Europe, between the Lower Rhine the system F6–F7, which may significantly offset right-laterally the Embayment and the North Sea, Geol. Soc. Am., 425, 193–224. Fault activity in the Dover Strait 541

Chadwick, R.A., Evans, D.J. & Holliday, D.W., 1993. The Maryport fault: Neilson, G., Musson, R.M.W.& Burton, P.W., 1984. The London Earthquake the post-Caledonian tectonic history of southern Britain in microcosm, of 1580, April 6, Eng. Geol., 20, 113–141. J. Geol. Soc., 150, 247–250. Ottemoller,¨ L., Baptie, B. & Smith, N.J.P., 2009. Source Parameters for the Clark, C.D., Gibbard, P.L. & Rose, J., 2004. Pleistocene glacial limits in 28 April 2007 Mw 4.0 Earthquake in Folkestone, United Kingdom, Bull. England, Scotland and Wales, in Developments in Quaternary Sciences, geol. Soc. Am., 99, 1853–1867. pp. 47–82, eds Ehlers, J. & Gibbard, P.L., Elsevier. Owen, H.G., 1975. The stratigraphy of the Gault and Upper Greensand Colbeaux, J.P., Somme, J. & Tuffreau, A., 1981. Tectonique quaternaire dans of the Weald, in Proceedings of the Geologists’ Association, Vol. 86, le Nord de la France: l’apport du gisement paleolithique´ de Biache-Saint- pp. 475–498. Vaast, Bulletin de l’Association franc¸aise pour l’etude´ du quaternaire, Radley, J.D. & Allen, P., 2012. The Wealden (non-marine Lower Creta- 18, 183–192. ceous) of the Weald Sub-basin, southern England, in Proceedings of the Collier, J.S, Gupta, S., Potter, G. & Palmer-Felgate, A., 2006. Using Geologists’ Association, Vol. 123, pp. 245–318. bathymetry to identify basin inversion structures on the English Chan- Reynaud, J-Y., Tessier, B., Auffret, J-P., Berne,´ S., De Batist, M., nel shelf, Geology, 34, 1001–1004. Marsset, T. & Walker, P.,2003. The offshore Quaternary sediment bodies De Bethune,´ P. & Bouckaert, L., 1968. Geological Map of Belgium of the English Channel and its Western approaches, J. Quater. Sci., 18, and Neighbouring Countries: Brussels, Geological survey of Belgium 361–371. Miscellaneous Geological Maps, scale 1:2 000 000, 1 sheet. Ruffell, A.H. & Wach, G.D., 1991. Sequence stratigraphic analysis of the Destombes, J.P., Shephard-Thorn, E.R., Redding, J.H. & Morzadeckerfourn, Aptian-Albian Lower Greensand in southern England, Mar. Petrol. Geol., M.T., 1975. Buried valley system in the Strait of Dover, Phil. Trans. R. 8, 341–353. Soc. A.: Math. Phys. Eng. Sci., 279, 243–256. Smith, A.J., 1985. A catastrophic origin for the palaeovalley system of the Downloaded from Everaerts, M. & Mansy, J.-L., 2001. Le filtrage des anomalies eastern English Channel, Mar. Geol., 64, 65–75. gravimetriques;´ une cle´ pour la comprehension´ des structures tectoniques Stein, S. & Mazzotti, S., 2007. Continental Intraplate Earthquakes: Science, du Boulonnais et de l’Artois (France), Bulletin de la Societ´ eG´ eologique´ Hazard and Policy Issues, Geol. Soc. Am., Spec. Paper, 425, 402. de France, 172, 267–274. Toucanne, S. et al., 2009. A 1.2 Ma record of glaciation and fluvial discharge Froidevaux, C., Paquin, C. & Souriau, M., 1980. Tectonic stresses in France: from the West European Atlantic margin, Quarter. Sci. Rev., 28, 2974– in situ Measurements with a flat jack, J. geophys. Res., 85, 6342–6346. 2981. http://gji.oxfordjournals.org/ Gibbard, P.L.,1995. The formation of the Strait of Dover, Geol. Soc., Lond., Underhill, J.R. & Paterson, S., 1998. Genesis of tectonic inversion structures: Spec. Publ., 96, 15–26. seismic evidence for the development of key structures along the Purbeck– Grunthal,¨ G., Wahlstrom,¨ R. & Stromeyer, D., 2009. The unified catalogue Isle of Wight disturbance, J. Geol. Soc., Lond., 155, 975–992. of earthquakes in central, northern, and northwestern Europe (CENEC)— Vandycke, S., Bergerat, F. & Dupuis, C., 1988. Paleo-contraintes´ a` la lim- updated and expanded to the last millennium, J. Seismol., 13, 517–541. ite Cretac´ e-Tertiaire´ dans le Bassin de Mons (Belgique). Implications Gupta, S., Collier, J., Palmer-Felgate, A. & Potter, G., 2007. Catastrophic cinematiques.´ Relations avec la zone de Cisaillement Nord-Artois, C. R. flooding origin of shelf valley systems in the English Channel, Nature, Academie´ de Sciences de Paris, 307(II), 303–309 448, 342–346. Van Vliet-Lanoe,¨ B., Mansy, J.L., Margerel, J.P., Vidier, J.P.,Lamarche, J. & Hamblin, R.J.O., Crosby, A., Balson, P.S., Jones, S.M., Chadwick, R.A., Everaerts, M., 1998. Le Pas de Calais un detroit´ cenozo´ ¨ıque a` ouverture Penn, I.E. & Athur, M.J., 1992. United Kingdom offshore regional report: multiple, C. R. Academie´ de Sciences de Paris, 326(II), 729–736. at North Dakota State University on May 20, 2015 the geology of the English Channel. HMSO for the British Geological Van Vliet-Lanoe,¨ B., Laurent, M., Bahain, J.L., Balescu, S., Falgueres,` C., Survey, London. Field, M., Hallegou´ et,¨ B. & Keen, D.H., 2000. Middle Pleistocene raised Hopson, P.M., Wilkinson, I.P. & Woods, M.A., 2010. A stratigraphi- beach anomalies in the English Channel: regional and global stratigraphic cal framework for the Lower Cretaceous of England. Research Report implications, J. Geodyn., 29, 15–41. RR/08/03. British Geological Survey, Keyworth. Van Vliet-Lanoe,¨ B. et al., 2002a. Cenozoic evolution of the Channel and James, J.W.C. et al., 2002. GEOSYNTH: A synthesis of the geology and southern North Sea areas (Western Europe). The neotectonic control on sediments of the Dover Strait and its hinterland. CR/02/078, CDROM, sedimentation and rivers, in Contributions to the Geology of Belgium and British Geological Survey. Northwest Europe: Proceedings of the First Geologica Belgica Interna- Lericolais, G., Auffret, J.-P.& Bourillet, J.-F.,2003. The Quaternary Channel tional Meeting, Leuven, Vol. 12, pp. 21–24, eds Degryse, P. & Sintubin, River: seismic stratigraphy of its palaeo-valleys and deeps, J. Quater. Sci., M., 11–15 September 2002. Aardkundige Mededelingen. 18, 245–260. Van Vliet-Lanoe,¨ B. et al., 2002b. Palaeogeographic evolution of North- Mansy, J.-L., Manby, G.M., Averbuch, O., Everaerts, M., Bergerat, F., Van western Europe during the upper Cenozoic, Geodiversitas, 24, 511– Vliet-Lanoe, B., Lamarche, J. & Vandycke, S., 2003. Dynamics and in- 541. version of the Mesozoic Basin of the Weald-Boulonnais area: role of Van Vliet-Lanoe,¨ B., Mansy, J.L., Henriet, J.-P., Laurent, M. & Vidier, J.P., basement reactivation, Tectonophysics, 373, 161–179. 2004. Une inversion tectonique cenozo´ ¨ıque par etapes´ : le Pas-de-Calais, Mellet, C., Hodgson, M., Plater, A., Mauz, B., Selby, I. & Lang, A., 2013. Bulletin Societ´ eg´ eologique´ Franc¸ais, 175-2, 175–195. Denudation of the continental shelf between Britain and France at the Warren, S.T. & Harris, C.S., 1996. An interpretation of the structural ge- glacial–interglacial timescale, Geomorphology, 203, 79–96. ology, in Engineering Geology of the Channel Tunnel, pp. 421–435, eds Melville, C.P., Levret, A., Alexandre, P., Lambert, J. & Vogt, J., 1996. His- Harris, C.S., Hart, M.B., Varley, P.M. & Warren, C.D., Thomas Telford torical seismicity of the Strait of Dover/Pas de Calais, Terra Nova, 8, Ltd. 626–647. Wells, D.L. & Coppersmith, K.J., 1994. New empirical relationships among Minguely, B., Averbuch, O., Patin, M., Rolin, D., Hanot, F. & Bergerat, magnitude, rupture length, rupture width, rupture area, and surface dis- F., 2010. Inversion tectonics at the northern margin of the Paris basin placement, Bull. geol. Soc. Am., 84, 974–1002. (northern France): new evidence from seismic profiles and boreholes Woods, M.A., Wilkinson, P. & Hopson, P.M., 1995. The stratigraphy of interpolation in the Artois area, Bulletin de la Societ´ eG´ eologique´ de the Gault Formation (Middle and Upper Albian) in the BGS Ariesey France, 181, 429–442. Borehole, Bedfordshire, in Proceedings of the Geologists’ Association, Mortimore, R.N., 2011. A chalk revolution: what have we done to the Chalk Vol. 106, pp. 271–280. of England?, in Proceedings of the Geologists’ Association, Vol. 122, Zoback, M.D. & Grollimund, B., 2001. Impact of deglaciation on present- pp. 232–297. day intraplate seismicity in eastern North America and Western Eu- Musson, R.M.W., 2004. A critical history of British earthquakes, Ann. rope, C. R. Academie´ de Sciences de Paris/Earth planet. Sci., 333, Geophys., 47, 597–609. 23–33. 542 D. Garc´ıa-Moreno et al.

SUPPORTING INFORMATION (http://gji.oxfordjournals.org/lookup/suppl/doi:10.1093/gji/ggv041 /-/DC1). Additional Supporting Information may be found in the online ver- sion of this paper: Please note: Oxford University Press is not responsible for the con- Table S1. Summary of the available bathymetric datasets, indicating tent or functionality of any supporting materials supplied by the the extent and the resolution of the DTMs. MBES: multibeam echo authors. Any queries (other than missing material) should be di- soundings; SBES: single-beam echo soundings. rected to the corresponding author for the article. Downloaded from http://gji.oxfordjournals.org/ at North Dakota State University on May 20, 2015